Jensen Huang, the CEO of Nvidia, is dealing with increased competition from bespoke chips.
According to a BofA analyst, worries over Nvidia have driven the stock’s valuation close to historically low levels, which means significant gains may be ahead.
The forward price-to-earnings multiple for the semiconductor manufacturer’s (NVDA) shares is currently around 25. Vivek Arya of BofA stated in a note on Monday that this is comparable to earlier recent “troughs” in October 2023 and July 2022.
When the stock’s multiple hit these levels in the past, the forward P/E “usually rebounded” to a range between 30 and 40 in the next three to six months, which is excellent news for Nvidia investors, according to Arya. He pointed out that this has resulted in a median forward price-to-earnings multiple of 37 for the last five years.
See also: As evidence of Big Tech’s dominance, Alphabet and Nvidia alone account for one-third of the S&P 500’s gains this year.
Furthermore, Arya noted that Nvidia shares are currently selling at their largest-ever discount to Broadcom (AVGO) shares—roughly 40%. Normally, Nvidia shares trade at a 10% discount.
Conceptually, this represents estimations that, when considering the second half of 2026 and the entirety of 2027, “already implicitly” shift at least 10 points of AI market share in favor of Broadcom, according to Arya.
Investors are increasingly discussing how the demand for graphics processing units from Nvidia and Advanced Micro Devices (AMD) would be affected by application-specific integrated circuits created by companies such as Broadcom.
Read more: Google worries cause Nvidia’s shares to plummet. Do investors fail to see the point?
After Google unveiled Gemini 3, which it claimed outperformed leading AI models from rivals like OpenAI and Anthropic, Alphabet (GOOGL) (GOOG) saw a spike in share price last month. The tech company claimed to have trained the AI model using its tensor processing units, which caused trepidation in a market already concerned about Nvidia’s ability to maintain its entire market share in a cutthroat sector. For over ten years, Google and Broadcom have collaborated on TPUs.
But according to Arya, clients will probably find it more difficult to switch over the course of the upcoming year due to Nvidia’s “leading allocation” from Taiwan Semiconductor Manufacturing Co. (TSM) (TW:2330), which also produces chips for Broadcom, and the limited supply of chips. Furthermore, whereas Nvidia’s chips have been utilized more extensively, Google’s own chips “have been proven only in Google’s data center,”
Nvidia will likely maintain its “dominant market share,” according to Morgan Stanley analyst Joseph Moore, who said previous concerns about the danger posed by ASICs “are becoming overstated.” Moore stated in a note on Monday, nevertheless, that it’s unclear what will be able to change the unfavorable opinion of investors.
Revenue from AI chips is expected to grow “slightly faster” for Broadcom and AMD than for Nvidia in 2026, according to Moore’s Morgan Stanley team. “But that mostly just reflects supply-chain limitations of a $205 [billion] run-rate revenue stream,” Moore added. It is anticipated that the supply of essential AI components will be limited through the following year.
More from MarketWatch: Elon Musk suggests two non-Tesla stocks to invest in AI’s future
According to Moore, “customers’ biggest anxiety” over the next 12 months is acquiring enough of Nvidia’s products, particularly with its upcoming Rubin AI platform.
“Of course, everyone wants an alternative, and those alternatives will have good economics for some applications,” Moore stated. “TPU is by all accounts a solid alternative, and one that has meaningfully contributed to several key models.”
Moore pointed out that Nvidia’s $51 billion in data-center sales in the most recent quarter is almost 14 times Google’s TPU revenue, and the $10 billion in sequential revenue increase that Nvidia recently recorded is around three times that overall TPU figure.
